| Abstract | Ciljevi istraživanja. Procijeniti povezanost ekspresije Ki-67 antigena i p16INK4a proteina i DNA ploidije s prisutnošću HPV-a visokog i niskog rizika i s patohistološkim nalazom CIN-a. Na temelju dobivenih rezultata utvrditi koji je od parametara nabolji pokazatelj prisustva CIN-a 2 i CIN-a 3 i utvrditi mogu li se istovremenim određivanjem i kombinacijom različitih biljega postići točniji rezultati u razdvajanju blage displazije (CIN 1) od umjerene i teške displazije (CIN 2 i CIN 3). Materijali i metode. Ispitivana skupina sastojala se od 111 uzoraka tkiva vrata maternice, od čega je bilo 30 CIN 1, 30 CIN 2, 31 CIN 3 i 20 uzoraka tkiva invazivnog karcinoma cerviksa. Kao kontrolna skupina uzeto je 15 uzoraka normalnog tkiva cerviksa. Na svim uzorcima bili su urađeni sljedeći postupci: patohistološka analiza, grupna tipizacija na HPV visokog i niskog rizika lančanom reakcijom polimeraze, imunohistokemijska analiza upotrebom protutijela na antigen Ki-67 i na protein p16INK4a, protočna citometrija i određivanje DNA ploidije. Rezultati. U skupini normalnog tkiva HPV visokog rizika nije bio prisutan, a bio je učestaliji što je promjena epitela bila teža, pa je u CIN 3 skupini i skupini invazivnog karcinoma učestalost HPV-a visokog rizika bila 90% (χ2 = 71,75; P < 0,001). Kod normalnog nalaza nije bilo prisustva p16INK4a, da bi učestalost njegovog prisutva rasla do 100% u CIN-u 3 i invazivnom karcinomu (χ2 = 70,06; P < 0,001). p16INK4a je bio povezan s HPV-om visokog rizika u 100% slučajeva, a nije bio prisutan u stanicama s HPV-om niskog rizika (χ2 = 97,76; P < 0,001). U aneuploidnim uzorcima je učestalost HPV-a visokog rizika bila 88,6%, statistički značajno češća u odnosu na 45% kod diploidnih uzoraka (χ2 = 20,79; P < 0,001). U skupini normalnog epitela svi su uzorci tkiva bili diploidni, u CIN-u 1 je bilo 10% aneuploidija, u CIN-u 2 16%, u CIN-u 3 41%, a u invazivnom je karcinomu bilo 70% aneuploidnih uzoraka (χ2 = 33,21; P < 0,001). Uzorci u kojima je bio prisutan HPV visokog rizika imali su značajno više Ki-67 pozitivnih stanica, a te su stanice bile češće u srednjoj i gornjoj trećini epitela (Kruskal-Wallis test, P < 0,001). Između skupine s HPV-om niskog rizika i HPV negativne skupine uzoraka nije bilo razlike u prisutnosti Ki-67 antigena. Kako je rastao stupanj displazije, tako je postotak Ki-67 pozitivnih stanica rastao u srednjoj i gornjoj trećini epitela (Kruskal-Wallis test, P < 0,001). Brojanjem Ki-67 pozitivnih stanica na 100 µm bazalne membrane i uz graničnu vrijednost za CIN 2 + CIN 3 od >15 stanica, u 61 uzorku umjerene i teške displazije bio je jedan lažno negativan nalaz (1,6%), a u 45 uzoraka s normalnim ili CIN 1 nalazom bio je jedan lažno pozitivan nalaz (2,2%). Računanjem postotka Ki-67 pozitivnih stanica u srednjoj i gornjoj trećini epitela zajedno i uz prag normale od 33%, jedan od 61 (1,6%) CIN 2 ili CIN 3 slučajeva bio je ispod granične vrijednosti, tj. lažno negativan. Također, samo je jedan od 45 (2,2%) normalnih ili CIN 1 slučajeva bio iznad te granične vrijednosti, odnosno lažno pozitivan. Mjereći na oba načina, to bi odgovaralo osjetljivosti i specifičnosti od 98% za prognozu CIN 2/CIN 3 promjena. Imunohistokemijsku ekspresiju Ki-67 antigena i p16INK4a proteina, HPV status i DNA ploidni status uvrstili smo u multivarijantnu analizu, u kojoj je patohistološka dijagnoza bila zavisna varijabla. Kada smo kao Ki-67 varijablu uzeli ukupan broj Ki-67 pozitivnih stanica na 100 µm bazalne membrane, samo je Ki-67 bio značajno povezan s dijagnozom uz 97,2% točno predviđenih nalaza (R2 = 0,967; P = 0,013). Izostavljanjem Ki-67 varijable iz logističke regresije, pojavila se ekspresija p16INK4a proteina kao varijabla statistički značajno povezana s patohistološkim nalazom (R2 = 0,470; P < 0,001; 82,1% točno predviđenih nalaza). Izrazivši bojenje na Ki-67 antigen kao postotak pozitivnih stanica u pojedinim trećinama cervikalnog epitela, pokazalo se da su jedino postotci Ki-67 pozitivnih stanica u srednjoj i gornjoj trećini epitela bili značajno povezani s patohistološkom dijagnozom (R2 = 0,811). Zaključak. Učestalost infekcije HPV-om visokog rizika raste s težinom histoloških promjena u cervikalnom epitelu. Infekcija HPV-om niskog rizika nije povezana s pojavom umjerene i teške displazije ili karcinoma cerviksa. p16INK4a je biomarker infekcije HPV-om visokog rizika i pomaže pri razlikovanju dobroćudnih reaktivnih lezija od displastičnih lezija. Određivanje DNA ploidnog statusa protočnom citometrijom metoda je koja pokazuje dobru specifičnost, ali nedovoljnu osjetljivost da bi se mogla primijeniti za otkrivanje displastičnih promjena. Imunohistokemijska analiza cervikalnog epitela brojanjem Ki-67 pozitivnih stanica na 100 µm bazalne membrane i u pojedinim trećinama cervikalnog epitela je vrlo osjetljiva i specifična metoda diferencijacije blage displazije (CIN 1) od umjerene i teške displazije (CIN 2 i CIN 3). Ujedno je to i najtočnija od istraživanih metoda u procjeni cervikalne intraepitelne neoplazije. Kombinacijom određivanja HPV statusa, DNA ploidnog statusa i imunohistokemijske ekspresije p16INK4a i Ki-67 nije moguće postići veću točnost predviđanja cervikalne intraepitelne neoplazije u odnosu na samu imunohistokemijsku analizu bojenja na Ki-67 antigen ili p16INK4a protein. |
| Abstract (english) | Objective. To correlate the expression od Ki-67 antigen and p16INK4a protein and DNA ploidy status with the presence of high-risk and low-risk HPV and with the pathohistological diagnosis. Depending on these results to evaluate which parameter is the best indicator of the presence of CIN 2 and CIN 3 and to analyze if more accurate results in distinction of mild dysplasia (CIN 1) from moderate and severe dysplasia (CIN 2 and CIN 3) can be accomplished by combining various biomarkers. Materials and methods. The study group consisted of 111 cervical tissue samples, including 30 CIN 1, 30 CIN 2, 31 CIN 3, and 20 samples of invasive squamous cell carcinoma. Fifteen samples of normal cervical tissue served as the control group. All samples were analyzed pathohistologically, by PCR for HPV group typing, by flow cytometry for DNA ploidy status and immunohistochemically using antibodies against Ki-67 antigen and p16INK4a protein. Results. In the group of normal samples there was no high-risk HPV positivity. The proportion of high-risk HPV positives grew gradually reaching 90% in the CIN 3 and invasive cancer groups, but there was no low-risk HPV finding among CIN 2 and CIN 3 cases (χ2 = 71.75; P < 0.001). p16INK4a was not present in normal cervical tissue and its occurence raised to 100% in CIN 3 and invasive cervical cancer (χ2 = 70.06; P < 0.001). All high-risk HPV cases and none of the low-risk HPV cases were p16INK4a positive (χ2 = 97.76; P < 0.001). In tissue samples with aneuploidy high risk HPV was present in 88.6% cases, which is significantly more frequent comparing with 45% in diploid samples (χ2 = 20.79; P < 0.001). In the normal tissue group all samples were diploid and aneuploidy was present in 10% of CIN 1 cases, 16% of CIN 2 cases, 41% of CIN 3 cases and 70% of invasive cancer cases (χ2 = 33.21; P < 0.001). High-risk HPV positive cases exhibited significantly more Ki-67 positive nuclei per 100 µm basal membrane, which were more frequent in the middle and in the upper third of the epithelium (Kruskal-Wallis test, P < 0.001). The differences in immunostaining between low-risk HPV and HPV negative samples were not significant. The percentage of Ki-67 positive cells rose in the middle and in the upper third of the epithelium with severity of the dysplasia (Kruskal-Wallis test, P < 0.001). Counting total number of Ki-67 positive nuclei per 100 µm basal membrane and using the cutoff value for CIN 2 + CIN 3 of more than 15 cells, one of 61 (1.6%) cases of moderate/severe dysplasia was false negative and one of 45 (2.2%) normal/CIN 1 cases was false positive. Calculating the percentage of Ki-67 positive cells in the middle third and in the upper third layer of the epithelium, with the cutoff value of more than 33% for moderate/severe dysplasia, one of 61 (1.6%) CIN 2/CIN 3 cases was missed by Ki-67 immunohistochemistry. Also, only one of 45 (2.2%) normal/CIN 1 cases was above the cutoff value, meaning false positive. This would correspond to sensitivity and specificity od 98% for CIN 2/CIN 3 using any of the two models of calcualtion. HPV status, DNA ploidy status, immunohistochemical expression of Ki-67 antigen and p16INK4a protein were included in a multivariant analysis with pathohistological diagnosis as dependent variable. With total number of Ki-67 positive nuclei per 100 µm basal membrane as Ki-67 variable, only Ki-67 staining was significantly linked to pathohistology with 97.2% of correct predictions (R2 = 0.967; P = 0.013). Omitting Ki-67 variable from logistic regression, p16INK4a expression became statistically significant with 82.1% of correctly predicted cases (R2 = 0.470; P < 0.001). Using the percentage of Ki-67 positive cells in three epithelial layers as three independent variables in multivariant analysis, only the percentages of Ki-67 positive cells in the middle third and in the upper third layer of the epithelium were significantly connected with pathohistological diagnosis (R2 = 0.811). Conclusion. The incidence of the infection with high-risk HPV raises according to the severity of histological changes in the cervical epithelium. Low-risk HPV infection is not associated with moderate or severe displasia, or cervical cancer. p16INK4a is a biomarker of high-risk HPV infection, and is useful in the distinction of benign reactive lesions from true displastic lesions. The determination of DNA ploidy status by flow citometry is a very specific, but not an enough sensitive one to be useful in the analysis of displastic changes in the cervical epithelium. Immunohistochemical analysis of Ki-67 staining in cervical epithelium by counting the total number of Ki-67 positive cells per 100 µm basal membrane and the percentage of these cells in three epithelial layers is a sensitive and specific method of differentiation between CIN 1 and CIN 2/CIN 3 grades. The combination of HPV status, DNA ploidy status and immunohistochemical expression of p16INK4a protein and Ki-67 antigen is not better in the prognosis of cervical intraepithelial neoplasia compared to the analysis of Ki-67 staining or p16INK4a expression. Ki-67 immunostaining is the most accurate of the analyzed methods in the evaluation of cervical intraepithelial neoplasia and can serve as a valuable adjunctive method for more accurate CIN grading in routine practice. |
